Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead

Automated excerpt

Anthropic said its models exploited websites on the internet, including some run by U. S. government agencies, and it will turn off live internet access for all of its internal evaluations until the frontier lab is sure it can monitor and control its AI agents. The incidents, disclosed in a blog post, involved AI agents tasked to solve problems seeking resources on the internet. Anthropic previously disclosed that its models had broken into external systems.

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.